Kismayo Semi-Quantitative Evaluation of Access And

KISMAYO SEMI-QUANTITATIVE EVALUATION OF ACCESS AND COVERAGE (SQUEAC) FINAL REPORT LOCATION: Kismayo, Somalia DATE OF INVESTIGATION: 8th December 2016 to 26th December 2016 TYPE OF INVESTIGATION: SQUEAC TYPE OF PROGRAMME: OTP for SAM IMPLEMENTING ORGANISATION: Skills Active Forward – UK (SAF – UK)

Acknowledgements

The assessment lead team wishes to express its gratitude to the following individuals or organizations for their efforts towards the success of the SQUEAC assessment in Kismayo:

• Ministry of Health (MoH) Jubaland for their support throughout the assessment and also

for granting authority for the assessment to be conducted

• UNICEF Somalia for funding the assessments

• AIMWG for their technical inputs and supports throughout the whole assessment period

• Himilo Foundation, Physicians Across Continents (PAC), International Committee on

Red Cross/Cresent (ICRC), Somalia Red Crescent Society (SRCS) and Somali Aid for

providing their inputs during field work

• The entire community in Kismayo who were welcoming and gave information freely and

to the best of their knowledge

• Finally but not least, Skills Active Forward – UK (SAF – UK) for their tireless inputs in

planning and coordinating the entire assessment and also mobilizing the community on

the assessment

CONTENTS Acknowledgements ...... ii

CONTENTS ...... iii

ABBREVIATIONS ...... vi

EXECUTIVE SUMMARY ...... vii

INTRODUCTION ...... 2

1.1 Justification ...... 3

1.2 Objectives of the Assessment ...... 3

1.3 Methodology ...... 3

STAGE 1 ...... 3

2.1 Quantitative data ...... 3

2.1.1 Admission trends ...... 4

2.1.2 % Requiring inpatient at admission ...... 6

2.1.3 Distance from treatment centre: ...... 6

2.1.4 Time to default: ...... 6

2.1.5 Length of Stay: ...... 7

2.1.6 Discharge Outcomes ...... 8

2.2 Qualitative data ...... 10

2.2.1 Concept map ...... 10

2.2.2 Barriers and Boosters ...... 10

STAGE 2 ...... 14

3.1 Methodology: ...... 15

3.2 Case definition: ...... 16

3.3 Prior formation ...... 17

3.3.1 Histogram prior: ...... 17

iii

3.3.2 Concept map ...... 17

3.3.3 Un-weighted barriers and boosters ...... 17

3.3.4 Weighted barriers & boosters: ...... 18

STAGE 3 ...... 18

4.1 Sample size & Precision: ...... 18

4.2 Quantitative sampling framework: ...... 19

4.2.1 Case finding methodology ...... 19

4.2.2 Quantitative data results: ...... 19

4.2.3 Reasons for SAM non-attendance ...... 20

RECOMMENDATIONS ...... 21

Annexes ...... 24

Annex 1: List of people trained during SQUEAC Assessment ...... 24

Annex 2: Training chronogram ...... 24

Annex 3: Kismayo OTP operational days ...... 25

Annex 4: Wide area survey quantitative data ...... 26

Annex 5: Health facilities Kismayo ...... 27

Annex 6: Concept map for Kismayo ...... 28

Annex 7: Kismayo Urban population estimates ...... 29

Annex 8: Map of Somalia showing position of Kismayo District ...... 30

List of Figures Figure 1: Trends in admissions in line with the seasonality ------4

Figure 2: Cumulative number of admissions by OTP site ------5

Figure 3: Oedema admissions by site (January – November 2016) ------5

Figure 4: The week that beneficiaries stopped attending OTP before recovery ------7

Figure 5: Length of stay for the cured discharges ------7

Figure 6: Trends in program discharge outcomes in line with seasonality ------8

Figure 7: BayesSQUEAC-calculator plot of the prior mode; prior α 9.9 prior β 10.9 ------18

Figure 8: Beta binomial conjugate analysis of single coverage estimate ------20

Figure 9: Caregiver reasons for not taking their malnourished children for OTP ------20

List of Tables Table 1: Comparison between OTP and SC admissions in September, October and November ...... 6

Table 2: Triangulated program boosters identified through the assessment ...... 10

Table 3: Triangulated program barriers identified through the assessment...... 12

Table 4: Stage 2 results of the small area survey ...... 16

Table 5: Analysis and interpretation of the small area survey results ...... 16

Table 6: Weighted and un-weighted barriers and boosters ...... 17

Table 7: Results of the wide area survey ...... 19

ABBREVIATIONS APD - Academy for Peace and Development ARC - American Refugee Committee AWD - Acute Watery Diarrhea CMN - Coverage Monitoring Network CHW - Community Health Worker FSNAU - Food Security and Nutrition Analysis Unit ICRC - International Committee of the Red Cross IDP - Internally Displaced Person LQAS - Lot Quality Assurance Sampling MoH - Ministry of Health MoP - Ministry of Planning MUAC - Mid Upper Arm Circumference NGO - Non-Governmental Organization OPD - Out-Patient Department OTP - Outpatient Therapeutic Program PAC - Physicians Across Continents RUTF - Ready to Use Therapeutic Food SAF UK - Skills Active Forward United Kingdom SAM - Severe Acute Malnutrition SC - Stabilization Center SQUEAC - Semi-Quantitative Evaluation of Access and Coverage SRCS - Somali Red Cross Society SWACEDA - Serve Women and Children Empowerment Development Agenda TSFP - Targeted Supplementary Feeding Program UNICEF - United Nations Children’s Fund W/H - Weight for Height WHO - World Health Organization

EXECUTIVE SUMMARY Kismayo district is in Lower Juba Region, South Central Somalia and has an estimated population of 376,293 (early December 2016) of which 8.2% (30,690) are in Internally Displaced Person (IDP) camps.1 The district is home to Kismayo town which is a port city and the third largest city in Somalia; it serves as the headquarters of the autonomous Jubaland State. The district borders Afmadow District to the North, Badhadhe District to the East, Jamaame District to the West and the Indian Ocean to the South. The major livelihood zones in Kismayo include pastoral and agro-pastoral in the rural areas and petty trading in the urban areas of Kismayo Town. The nutrition situation in Kismayo urban is estimated as 8.8%2, which is classified as “alert” based on the World Health Organization classification of malnutrition, while the situation in the IDPs is estimated as 14.5%3, which is classified as “serious”. The 10 sites included in the assessment were under one implementing partner of which three sites were in the IDP camps. In addition, there are five sites, which offer the Targeted Supplementary Feeding Program (TSFP) and are run by Himillo Foundation and there are two other sites run by International Committee of the Red Cross (ICRC) and Physicians Across Continents (PAC) that offer inpatient services for complicated Severe Acute Malnutrition (SAM) cases. This assessment took place between 8th and 26th December 2016 and was led by a Semi-Quantitative Evaluation of Access and Coverage (SQUEAC) specialist trained by Coverage Monitoring Network (CMN). The program for the 10 Outpatient Therapeutic Program (OTP) sites that was included in the assessment commenced in January 2016 with some sites being operational as of April 2016. Oxfam was among the Non-Governmental Organizations (NGOs) that implemented OTP previously before Skills Active Forward United Kingdom (SAF UK) came into Kismayo District. Currently, the OTP implementing organizations include SAF-UK, PAC, Somali Aid and ICRC. The single coverage estimate realized by use of the 10 OTP sites in Kismayo district that availed their program data is 46.8% (40.8% - 53.1%). The coverage is much lower than the Sphere standard for an urban set up (70%).

Main barriers Explanation Ready-to-UseTherapeutic RUTF is shared in the homes and exchanged for food stuff at Food (RUTF) as food shops where it is sold to willing customers OTP not integrated with The OTP sites were perceived to be generally near and deliver other medical treatment only OTP services. However, the nearest medical treatment services centre is the District hospital, which the community said was very far to walk with a sick child Weak OTP-SFP interface OTP and TSFP are offered by separate partners with separate operating procedures. Respondents indicated that the TSFP sites were far. Some beneficiaries would therefore not be able to continue with treatment of malnutrition after OTP. Late treatment seeking Caregivers first visit community traditional healers, religious leaders (for prayers) and OTP comes as a last resort. Influx of families from IDP repatriation from Kenya and others coming in from the

1 MoP Kismayo 2016 IDP population estimates; Village elders and program coordinator for host estimates 2 Somalia FSNAU 2015/2016 Post Deyr Report 3Somalia Food Security and Nutrition Analysis, Post Gu 2016. Technical Series Report No. VII. 69, October 2016 vii

neighboring areas neighboring areas going through military operations Limited information Partners with programs offering nutrition treatment services sharing among partners within the same community have differential information on services offered and any nearby sites for referral of beneficiaries Main Boosters Explanation Positive opinion of OTP Respondents said child malnutrition has somewhat reduced due to the OTP services. Active screening and Screening is often done door to door with physical referrals of follow up of beneficiaries those who take time to reach the OTP site. Strong community links Children in OTP are known by the community volunteers, community leaders of the sections and the Community Health Workers. Community can identify Respondents mentioned oedema (barar), thinness, loss of some signs and buttocks, stretching skin and fatigue relating to nafaqadaro (local symptoms of term relating to under-nutrition). malnutrition

Several recommendations were drawn based on the findings of the assessment. These included: sensitization meetings in the community on malnutrition and its treatment by Ministry of Health with village elders, other community groups, and influential individuals in the community; beneficiaries to bring back the empty sachets during their visits before being given more ration; sensitization of caregiver groups on income generating activities in order to counter the sale of RUTF; admission into the program to consider all the admission criteria including Mid Upper Arm Circumference; and caregivers to be informed of the admission procedures without promise of OTP admission. Other recommendations included: OTP sites to scale up the services offered to include some medical treatment services; increase number of TSFP sites; have agenda for program information sharing with partners during coordination meetings; increase RUTF buffer supplies to ensure needs remain met even when there are challenges with transportation and delivery; and increase active case finding during months of population influx.

viii

INTRODUCTION Kismayo is a district in Lower Juba Region, South Central Somalia and is generally flat and sandy. The area had been experiencing prolonged drought due to failed rains such that the assessment was conducted under hot and humid weather conditions.

The population size in the town has increased drastically within the assessment year (2016) due to influx of families and IDPs from neighboring areas4. Currently the population of Kismayo town (urban and IDP) is estimated at 376, 293 (early December 2016) of which 8.2% (30,690) are in IDP camps5. The under-5 population is estimated at 20%. The area is an Islamic state with predominantly Somali speaking residents. The major livelihood zones in Kismayo include pastoral and agro-pastoral in the rural areas and petty trading in the urban areas on Kismayo Town. Residents of Kismayo area receive most of their food through imports that arrive by road or through the harbor at the port of Kismayo6. Fruits can be found in the Somali communities (mainly the Bantu) living along the riverine. Kismayo urban is more stable in terms of security as evidenced by the availability of medical treatment services and NGO presence7. Rural Kismayo, as in many other neighboring areas, is controlled by militia. Some inhabitants in these areas find ways of moving into more secure areas (like Kismayo urban) where they can find medical treatment and food8. The nutrition situation in Kismayo urban was estimated as 8.8%9, which is classified as “alert” based on the WHO classification of malnutrition, while the situation in the IDPs is estimated as 14.5%10 that is classified as “serious”. The actual number of operational OTP sites serving the population in Kismayo town could not be established due to the challenge of limited information sharing among partners. Through the assessment, however it was established that there are at least 10 confirmed operational OTP sites as well as 2 operational SCs implemented by PAC and ICRC to treat SAM cases. The 10 OTP sites offer no medical treatment services apart from protocols for management of uncomplicated SAM cases. Before children 6-59 months are admitted to OTP, anthropometric measurements (W/H, MUAC and checking for oedema), medical examination for complications and an appetite test are conducted to direct admission or referral11. OTP takes in those children that either have a MUAC < 11.5cm, W/F Z score of < -3, Oedema + or ++12. The children are to visit the OTP site weekly or biweekly and be provided with a take-home ration of RUTF (dosage according to current weight) and routine medication as guided by the protocol until they are “cured” (either W/H > -2z score, MUAC > 12.5cm or no oedema for two consecutive visits).13 The 10 sites have used W/H since April 2016 as the preferred admission criteria recommended by UNICEF14. Discharge criteria in the 10 sites that

4 Interviews with program staff, OTP team, caregivers 5 MoP 2016 for IDP population and village elders & program coordinator for host population 6 Interviews with Program staff, SQUEAC enumerators and community members 7 SC in charge, program staff 8 Interviews with SQUEAC enumerators, program staff, caregivers within the community 9 FSNAU Post Deyr Report 2015/2016 10Somalia Food Security and Nutrition Analysis, Post Gu 2016. Technical Series Report No. VII. 69, October 2016 11 UNICEF (2010). Somali Guidelines for Management of Acute Malnutrition 12 UNICEF (2010). Somali Guidelines for Management of Acute Malnutrition 13 UNICEF (2010). Somali Guidelines for Management of Acute Malnutrition 14 Interviews with program staff, OTP team leaders and OTP supervisor 2

were involved in the assessment was therefore W/H > -3 Z score after which discharge to their nearest TSFP site is to be done15. 1.1 Justification SAF UK has been operating the outpatient therapeutic program in Kismayo District since January 2016. As a result, there was need to assess the program with an aim of determining the major barriers and boosters, which influence the program with an ultimate goal of improving service delivery in the community it serves. 1.2 Objectives of the Assessment The assessment was guided by the following specific objectives: 1. To identify factors (Barriers and Boosters) affecting the uptake of OTP services in Kismayo District 2. To establish the overall coverage estimate for the OTP program in Kismayo District 3. To provide action plan to improve acceptance and coverage of OTP in Kismayo district 4. To enhance capacity of MoH and other Program Staff from partners in using SQUEAC methodology to assess program coverage in Kismayo District 1.3 Methodology The SQUEAC Methodology was conducted by design in three phases namely, Stage 1, Stage 2 and Stage 3. Stage 1 involved identifying areas of low and high coverage as well as reasons for coverage failure using routine program data, any other existing data and qualitative data (mainly interviews and observation). Stage 2 involved confirming the location of areas of high and low coverage and the reasons for coverage failure identified in stage 1. This was done using a small- area survey. Stage 3 involved provision of an overall estimate of program coverage using Bayesian techniques. Challenges and limitations Conflicting information was given on number of operational sites and services offered among partners during the assessment. Information from triangulated sources was therefore given first consideration. Unfortunately, only one partner shared their routine OTP data therefore the coverage barriers and boosters may have been slightly different for the remaining sites. Nevertheless, weighing of the barriers and boosters took note of inputs from enumerators who worked with representative OTP partners for a more representative situation analysis. Qualitative data collection techniques (of following information leads) took time to be mastered. To address this, daily debriefing sessions with the team were carried out for clarifications, while each pair of enumerators was asked to take turns conducting the different interviews while working as team. STAGE 1 2.1 Quantitative data Data on admissions by month and OTP site (Figure 1&2), oedema admissions by OTP site (Figure 3), time to default (Figure 4), length of stay (Figure 5), discharge outcomes by month (Figure 6), and percentage of beneficiaries requiring inpatient care at admission (Table 1) was collected.

15 Interviews with Program staff, caregivers of some OTP cured beneficiaries, OTP site supervisor and team leaders 3

2.1.1 Admission trends The average admission per month (January to November 2016) for all the OTP sites was 371 SAM cases. A trend that is distinctive of a program in its start up phase was evident in Kismayo where the year began with high admissions that quickly dropped. All the same, the admissions seemed to be mainly influenced by high disease occurrence, periods of drought and Influx of populations from neighboring areas. More specifically, March, May, August and November had high numbers of admissions fueled by high numbers of incoming IDPs from Kenya (returnees) and Gosha (neighboring areas militia controlled)16. In the same period, there were high AWD, measles and malaria cases in the community. Measles campaign began with several other IDPs arriving into the area (Figure 1).

Figure 1: Trends in admissions in line with the seasonality

Season Jan Feb Mar April May Jun Jul Aug Sept Oct Nov Diseases High AWD AWD, measles and Malaria High measles

Climate Hot and Floods Drought Floods Drought dry Food price Low food prices High food prices Migration IDP returnees Military operation outside Military Kismayo and returnees operation Program RUTF New RUTF stock OTP stock out sites out Labor Cultivation, Constructions, Offloading ships

Dalxiska camp has a fixed OTP site and, being the largest IDP, had the highest cumulative admissions. It started as an outreach site then changed into a fixed site (by April 2016) due to the large number of malnourished children that would turn up. Dabyanbo site (fixed) is part of the host community with just a few IDPs in addition to it being one of the initial sites of the 10 sites

16 Interviews with village elders, program staff and OTP team 4

that were assessed. The sites with few cumulative admissions began in April 2016 and with smaller population sizes (Figure 2).

Figure 2: Cumulative number of admissions by OTP site

Figure 3: Oedema admissions by site (January – November 2016) Five percent (206) of the cumulative admissions in the assessed sites were oedema admissions (Figure 3). Sites within highly populated zones (Annex 7) such as Hilac in Shaqalaaha zone and Dalxiska in Farjano zone, were observed to be congested in housing and were seen to have more 5

cases of Oedema. Interviews with enumerators, program staff and OTP site supervisor revealed that Alanley’s Dowdhanan area was at a riverine border with inhabitants who fish for food (protein). This could explain why the OTP site has not had Oedema cases since it became operational in April 2016.

2.1.2 % Requiring inpatient at admission Since April 2016, the 10 sites that were assessed had been using their donor preferred admission criteria, weight for height Z-score to mean there have been no MUAC admissions since then.

Table 1: Comparison between OTP and SC admissions in September, October and November Admissions Conclusion Month OTP SC % Average % Sept 267 369 58.0% The late treatment seeking is 53.8% Oct 283 382 57.4% high. Nov 457 389 46.0%

Proportion of children with SAM who required inpatient care at the time of admission was taken to be proxy for timeliness of admission (treatment seeking). Alternatively, a comparison was done between the number of children admitted into the SC and those admitted into the OTP in a span of 3 months (September, October and November). This was done as a proxy for the timeliness of admissions within the OTP sites in Kismayo being that more children had been admitted into SC, as new SAM with complications, in comparison to inter-program referrals17. Children requiring inpatient care should not exceed 5% in an already established program site18 (Table 1).

2.1.3 Distance from treatment centre: To caregivers, perception of distance differed depending on personal preferences. For 10 OTP sites that were assessed, all were within the various zones in the area where the farthest village from their nearest OTP site was about 15 to 20 minutes walk. Some said it was very near while others viewed it as far. It was however evident that all the groups and individuals interviewed felt that their only public hospital (the district hospital) was very far to walk with a sick baby. When some children would get ill, the caregivers would sometimes not be able to leave their other children to get medical treatment services. Soon after, most of those children would deteriorate to acute malnutrition.

2.1.4 Time to default: Time to default was taken for the 3 months; September, October and November, preceding the assessment. All the 10 sites assessed had late defaulters who left the program at 4 months and beyond (Figure 4).

17 Interview with SC in-charge 18 Myatt, Mark et al. 2012. Semi-Quantitative Evaluation of Access and Coverage SQUEAC)/ Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Technical Reference. Washington, DC: FHI 360/FANTA. 6

Figure 4: The week that beneficiaries stopped attending OTP before recovery Interviews conducted on the OTP supervisor, OTP team leaders the CHWs revealed that most of the defaulters are in families that migrate in and out of the Kismayo area. Actual reasons from caregivers of the defaulters were thus not established during this assessment.

2.1.5 Length of Stay: Length of stay was taken for children dischared as cured in the 3 months (September, October and November) preceding the assessment.The median length of stay was 8 weeks to signify that most children are discharged as cured by the time they are 8 weeks old in OTP.This means that children stay in the program for a short time19 (Figure 5).

Figure 5: Length of stay for the cured discharges

19 Myatt, Mark et al. 2012. Semi-Quantitative Evaluation of Access and Coverage SQUEAC)/ Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Technical Reference. Washington, DC: FHI 360/FANTA. 7

2.1.6 Discharge Outcomes Death occurrence between February (17.1%) and April (17.0%) was higher than the Sphere (10%) for the management of SAM. This could have been attributed to the then acute watery diarrhea outbreak that affected both children and their caregivers in Kismayo. As well, defaulting almost reached the limit Sphere (15%) in the same period (March 12.7%; April 12.0%) due to a number of possible reasons. Those months were hot to walk for distances with a sick child20 as floods along the Riverine would have strained accessibility of the OTP sites by the caregivers. RUTF stock-outs was likely to have resulted in absenteeism21 and later defaulting; labor demands on the caregivers where they would go to work and alongside their children22 increasing their chances of defaulting. As well, the AWD outbreak that affected families at large might have resulted in some caregiver admission into inpatient care and children missed the OTP visits. As revealed through interviews on team leaders, Program staff, OTP supervisor and some caregivers of uncovered children, a sick mother is not in a physical position to attend OTP and her child will therefore be a defaulter in the end.

Season Jan Feb Mar April May Jun Jul Aug Sept Oct Nov Diseases High AWD AWD, measles and Malaria High measles

20 Interviews with program staff, OTP supervisor and village elders 21 Interviews with CHWs, Team leaders, program staff 22 Interviews with program staff and enumerators 8

The cure rate between January and April was consistently below the Sphere standard (75%), which is consistent with high defaulting and deaths that could hinder recovery. The cure rate

between May and August tallies with the scale up of OTP sites to the 10 (that took part in the assessment) that increased number of cases treated and discharged successfully. The downward spike in the month of September tallied with the measles outbreak that might have caused complication of SAM and the month had limited food availability due to high food prices. When food is not easily accessible, RUTF given to beneficiaries is likely to be shared within the household or exchanged for other foodstuff23 (Figure 6). 2.2 Qualitative data For individual interview guide that was necessitated in the process of stage 1, sampling was done on 30% of the 10 operational sites whose program data was assessed. Where more information was needed to clear the questions that arose, interview guides targeting the individuals best fit to offer the clarification were utilized to no new information could be gathered. 2.2.1 Concept map This was drawn using the Epigram software to elaborate how the identified barriers and boosters relate in the context of Kismayo (Annex 6). 2.2.2 Barriers and Boosters The triangulated barriers and boosters were classified using the information collected from different sources and methods (Table 2 & 3). Table 2: Triangulated program boosters identified through the assessment Boosters Explanation and analysis Sources 1 Positive • Community members, say OTP Improves child health through the treatment and use of OTP team leaders (nurses opinion of Plumpy nut in charge of OTP sites), OTP • Chairwomen (female village elders) said the situation of nafaqadaro is not the same CHWs, OTP supervisor, way it was before, things are getting better however there is still much to do. chairwomen, Caregiver of • Beneficiary said she learnt about the program from a neighbor who informed her that beneficiary, Father of her child had been like hers and went to the OTP and he got better and that hers would children under 5years, too Sheikhs, Observation • When the Sheikh received a child who had nafaqadaro, he would say a prayer and then refer to the nearest OTP site or OPD for further assistance. 2 Active • Screening is done from door to door OTP supervisor, CHWs, screening and • Once admitted the CHWs and volunteers visit some of the beneficiary homes to verify caregiver of beneficiaries, follow up of compliance of use of RUTF Sheikh, Fathers of children beneficiaries • Beneficiaries who fail to attend OTP visits are followed up after the first day of failing under 5 years, OTP team

23 Interviews with caregivers of beneficiaries and those already OTP occurred, OTP supervisor, Shopkeepers and team leaders. 10

to attend. Some program staff stated that defaulters are attached to CHWs for follow leaders, program staff up. CHWs stated that they ask for the beneficiary contact for follow up purposes in case she/he defaults. 3 Screening at • All children who visit to the OTP site are taken MUAC, Height and weight OTP supervisor, CHWs, OTP site measurements, and checked for presence of Oedema OTP team leaders, program staff, observation 4 Strong • Children admitted into OTP are known by name by the community volunteers of the CHWs, area OTP community blocks, the community leaders of the sections and the CHWs who referred them. supervisor, Community links • Combined meetings (chaired by the OTP supervisor) are held for all Kismayu sites volunteers (chairwomen) every Thursday to look into challenges and share updates on nutrition and health. • Every month community leaders are tasked (by the OTP site supervisor) to mobilize all mothers in their areas to meet up at OTP site for health and nutrition education. • Some parents when children are sick will look for their community leader who will then connect them with their nearest OTP site. 5 Follow up on • CHWs ensure referrals reach the OTP site. Chair-women sometimes take the children Supervisor, chairwomen, community to their nearby sites themselves. program staff, CHWs referrals to • Referrals that were late to visit OTP site were escorted by the CHWs and sometimes OTP sites the community women to the OTP site. 6 OTP teams • The OTP teams (supervisor, team leaders, screeners, CHWs) received 5-day training, Observation, team leaders, are competent and on-job training on management of acute malnutrition with emphasis on SAM, and program manager, OTP public health promotion for community based workers. Supervisor, CHWs • The CHWs interviewed were able to identify the left upper arm as the point of measurement and are able to demonstrate the MUAC. 7 Referrals • Referral slips are used when children are referred to or from OTP, SFP or SC. Nutrition program between • Some beneficiaries are physically taken to the site of the next program for handover. coordinator, OTP site programs supervisor, OTP team leaders 8 Community • Mentioned signs edema, distended stomach, thinness, loss of buttocks, constipation, Chair-women, Sheikh, can identify lethargic, dehydration (stretching skin), white or brown hair that falls out Fathers of under 5, some signs caregivers of under 5. and symptoms of malnutrition (nafaqadaro)

Table 3: Triangulated program barriers identified through the assessment. Barriers Explanation and analysis Sources 1 RUTF as • The caregivers refer to Plumpy nut as buscut loos (peanut biscuit), buscut bajiq (soft biscuit). Program staff, food • CHWs, caregivers and some program staff were observed to use the term buscut to refer to CHWs, CHWs, Plumpy nut. CHWs stated that when mothers don’t attend their distribution day, the CHWs call supervisor, them and ask them to come for buscut loos caregiver of • Program attest to some caregivers saying that they visit the OTP site to see if their child is relapse qualified to get buscut (Plumpy nut) beneficiary, • Some mothers were reported to bring one of their children for follow up treatment visit with their observation, other children demanding admission into the program. team leader, • Shopkeepers said the caregivers bring the sachets of plumpynut to exchange for food items such Shopkeeper as tea leaves, milk and sugar after which they are sold (5 USD24 per satchet) to community. • When one child is given the ration, mothers are said to distribute it amongst all her young children, as she may not have enough food for them. 2 Community • Mothers referred by CHWs by MUAC (<11.5cm) reach the OTP site where confirmation of Team leaders, screening MUAC is done, weight and height is taken plus checking for edema. In the end, most children are CHWs, by MUAC only admitted when their W/H Z-score is found to be < -3SD Z score. Community while • Mothers of hopeful beneficiaries become disappointed and angry and sometimes use harsh words volunteers, admission on the OTP team for the inconsistency in the screening admission. Next time they asked to visit fathers in by W/H Z the site, they become hesitant. community, scores only • Community volunteers sometimes will get information that a beneficiary was sent back home wide area survey because her weight was not fitting criteria 3 OTP sites • The only district hospital in Kismayu area is about 1 hour from the villages and considered very OTP site not far by all. Community volunteers (chair-women) say there is no ambulance to help even in the supervisor, integrated case of medical emergencies. Observation, with other • They felt that OTP site are considered close as within 15 minutes from farthest village and within caregivers of medical the different camps, even so, they cannot get medical treatment there. The community members under 5s, services proposed that the OTP should expand and include the medical component even if just for OPD to chairwomen be able to treat the illnesses as they arise within community in time before deterioration into malnutrition. 4 Weak OTP- • Some partners offer OTP services only, some SFP services only while others have both program Program staff, SFP services. Each partner has at times different strategies and policies that sometimes will limit Caregiver of interface admission of OTP referrals. beneficiary,

24 The community mentioned between 3000 to 5000 Somali Shillings per Sachet which is approximated as 5 USD (using 1 USD ~ 600 Somali Shillings) 12

• The community volunteers and caregiver of a then OTP referral beneficiary also say the TSFP chairwomen ( sites known to them were quite far (15-20 minutes’ walk) from the Kismayu area to walk with a community sick child under the sun. volunteers) • Recent OTP referral (a Bantu caregiver of relapsed beneficiary) said she went to a TSFP site and was told that since she came from a different area (Dalxiska IDP camp) that she could not be assisted. She went back home and baby relapsed as could only feed her children on the little she got. This is even as some referrals from other camps were admitted as noted by community volunteers. 5 Limited • Partners with programs offering nutrition treatment services within the same community have Program information differential information on services offered, if there are any nearby sites to refer beneficiaries to manager, sharing them. program staff, among • Information on partners OTP services offered is not shared on the coordination platform program partners • The different organizations running SC, OTP and TSFP programs seem not to know where and coordinator, when different sites operate to aid in follow up of transfers, and know who is in what program. OTP Supervisor • Handover of beneficiaries to their home’s nearest OTP in Kismayu district, from SFP, only to find that it is a mobile site and do not work on that day. • Different criteria for programs are used in different organizations thus challenging some referrals • Once referral forms or letters are written for many beneficiaries, little follow up of receipt is done 6 Stigma • Some caregivers will keep their children away from neighbors’ children who are said or thought to Community have nafaqadaro as they tend to have diarrhea and are sickly. volunteers, • Some fathers as seen by some community volunteers and OTP team members will tend to divorce fathers of under their wives claiming that the mother did not take care of their children as seen in the evident 5, OTP Nafaqadaro. supervisor • They refuse their children to be taken to the OTP even after referral for fear of contracting skin infections due to sharing of the weights pants. 7 Late • Mothers are said to take their time to seek medical attention when it comes to their children’s OTP Team treatment illnesses. The OTP team leaders said caregivers visit the community traditional healers, religious leaders, seeking leaders for prayers, pharmacies, which in most cases the last resort being the OTP site, therefore Caregivers, SC cases. observational • About 54% of all SAM cases treated between September and November 2016 in the assessment study, CHW period required inpatient care in SC. This is much higher than the recommended 5% threshold. • CHWs stated that some mothers stay with malnourished and sick children in the house maybe due to her looking after her other children too until the baby gets diarrhea then they take to hospital. They therefore develop malnutrition quickly afterwards. • Some caregivers believe that malnutrition (nafaqadaro) is inborn and would not get to the OTP

site soon after they were referred 8 Stock-out • All the 10 sites that took part in the assessment experienced stock out in March 2016 and Oct- Program staff, of RUTF Nov 2016 where RUTF was out of stock for about 3 weeks each time. During those weeks, the OTP team OTP team leaders stated to there were “a lot of problems” especially because they had to try leaders, CHWs convincing some beneficiaries who would keep asking about them. 9 Influx of • The communities that are undergoing military operations and natural calamities (drought, floods) Enumerators, families are seeking refuge in the camps within Kismayo. Ongoing repatriation from Kenya has seen large program staff, and IDPs numbers of IDPs being accommodated by their relatives within Kismayu sharing resources and chairwomen, from amenities in tiny spaces. site supervisor, neighboring • Most of them come in already malnourished and in high numbers, therefore the camps are CHWs areas gradually becoming congested. Most of them also do not have food or economic capabilities for survival therefore deterioration of health. 10 Defaulting • Community volunteers were concerned that there seemed to be high defaulting due to the mothers Team leaders, looking for a means of income. Others move back to their families in the, neighboring areas, when community they are notified about their families needs. volunteers, CHWs, OTP supervisor

STAGE 2 A 2-part hypothesis on coverage was formulated from the information gathered through stage one. Part 1 read, “Coverage of OTP is high in IDP Camps” while the other part of the hypothesis stated, “Coverage of OTP is low in the host community”. The hypothesis was generated due to the following reasons: - Most of the admissions to OTP were from the site within the IDP camps - Sphere standard for program coverage in IDP camps setting is 90%25 - Community volunteers (also referred to as female village elders or chairwomen) in the IDP camps knew in more detail which households had children were sick, than their counterparts in host community. - IDP camps had houses very close to each other which was thought to enhance peer information sharing A Small area survey was conducted to test the hypothesis in the general population without regard for the spatial disposition.

25 Sphere Project. (2011). The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response (3rd ed.). Practical Action Publishing. Retrieved from http://www.sphereproject.org/handbook-download 14

3.1 Methodology: The largest IDP camp was selected, from which 2 smaller camps within that larger camp were randomly sampled for the small area survey. One zone of 4 host community zones was selected randomly then 2 villages in the selected zone were selected randomly for

hypothesis testing. Community volunteers (village elders/chairwomen) in-charge of the households was utilized to introduce the teams to the households and start the active and adaptive process. Once a child was screened using MUAC tape, the caregiver was asked if she knew of a child who was like her child now. The team would then follow the lead to that household. Sampling continued until there were possibly no other children who had SAM in the camps or villages. 3.2 Case definition: During the Small area survey, the term nafaqadaro was used in case finding. The term was established in stage 1 as the most common expression used by the community to mean a child who was very thin, usually due to lack of food. Cases were defined as either SAM in OTP (covered cases), SAM not in OTP (uncovered cases), Recovering SAM in program (covered cases in either OTP or SFP) and Recovering SAM not in program (Table 4).

Table 4: Stage 2 results of the small area survey Area Village/ SAM in OTP SAM NOT in Recovering Recovering SAM camp (MUAC OTP (MUAC < SAM in NOT in program name <11.5cm 11.5cm program MUAC ≥11.5cm or oedema) or oedema) MUAC ≥11.5cm Host Hilac 2 13 1 19 community Horsed 2 2 4 15 IDP camps Najax 5 3 5 7 Tawakal 5 14 13 40

The simplified LQAS decision rule d=n × p/100 was used in analysis of the hypothesis test. The percentage Sphere coverage standards of at least 70% in the urban setting and at least 90% in the IDP setting were used in the interpretation of the results. Coverage was to be labeled “high” if the covered children (SAM in OTP) were higher than the decision rule (d) and “low” if the covered children were less than the decision rule (d) (Table 5). Table 5: Analysis and interpretation of the small area survey results Villages in Host LQAS decision rule Conclusion community Coverage target 70% Since the number of covered cases (4) was lower Sample total (SAM cases) 19 than the decision rule (17), The hypothesis that Decision rule “OTP coverage is low in villages within host community” was accepted. This meant that

Number of SAM covered 4 coverage within the host community was less than70%. Camps within larger IDP LQAS decision rule Conclusion camp Coverage target 90% Since the number of covered cases (4) was lower Sample total (SAM cases) 27 than the decision rule (17), the hypothesis “OTP LQAS decision rule coverage is high in the IDP camp” was rejected. It also meant that coverage in the IDP was less Number of SAM covered 10 than 90%

A simple structured interview guide was utilized on the caregivers who had children presenting with SAM and were not in the program. This was done to gain their reasons for non-attendance and that were compiled together with the responses received through stage 3 (Figure 8).

3.3 Prior formation Four methods were used in formulation of the prior.

3.3.1 Histogram prior: A smooth curve showing the believed coverage was drawn using the prior knowledge of the barriers and boosters as derived through stage 1 and 2. Through stage 1, competent team, strong community It was thought to be likely that the coverage would lie between 40% and 60%, with the histogram prior at 45.0%

3.3.2 Concept map The arrows that represented a negative effect were counted as barriers, while those that symbolized positive effect as boosters. Twelve arrows for boosters were added from 0 while 14 arrows for barriers were subtracted from 100%. After division of the result by two, a prior of 49.0% was found (Annex 6). (0+12) + (100 - 14) = 98; 98/2 = 49.0%.

3.3.3 Un-weighted barriers and boosters A list of well-triangulated barriers and boosters was taken to have equal effect on coverage where each of them was given the maximum effect of 5%. There were 10 barriers and 8 boosters that added up to 50% and 40% respectively (Table 6). The percentage boosters were added to the 0, the total barriers were subtracted from 100%, the result was divided by 2 to give a prior of 45.0% (Table 6). (0+40) + (100 - 50) = 90; 90/2= 45.0%

Table 6: Weighted and un-weighted barriers and boosters Barriers Weighted Un- weighted RUTF as food 5% 5% Community screening by MUAC while admissions by W/H z-scores 2% 5% OTP not integrated with other medical treatment services 5% 5% Weak OTP-SFP interface 4% 5% Limited information sharing among partners 2% 5% Stigma 1% 5% Late treatment seeking 4% 5% RUTF stock-out 3% 5% Influx of families from neighboring areas 4% 5% Defaulting 3% 5% Total 33% 50% Boosters Weighted Un- weighted Positive opinion of OTP 5% 5% Active screening and follow up of beneficiaries 5% 5% Screening at OTP site 3% 5% Strong community links 5% 5% 17

Follow up of OTP community referrals 3% 5% Competent OTP teams 4% 5% Referrals between programs 4% 5% Community can Identify some signs and symptoms of malnutrition 5% 5% Total 34% 40%

3.3.4 Weighted barriers & boosters: Barriers or boosters thought to have more effect on coverage were assigned higher percentage where 5% was the maximum and 1% was the minimum. The barriers gave 33% while boosters 34% (Table 6). Barriers were subtracted from maximum coverage (100%) while the boosters were added to minimum coverage (0). The sum of the result was then divided by 2 to give a prior of 50.5% (Table 6). (0+ 34) + (100 - 33) = 101; 101/2 = 50.5%

A calculation of the prior mode was done and a prior plot using the BayesSQUEAC calculator version 3.01 was formulated (Figure 6).

(45.0+ 50.5+ 45.0+ 49.0)/4= 47.4%

Figure 7: BayesSQUEAC-calculator plot of the prior mode; prior α 9.9 prior β 10.9 STAGE 3 4.1 Sample size & Precision: A minimum sample size of 77 children was calculated using the following formula:

Where the shape parameters were Prior alpha (α) was 9.9 while the prior beta (β) was 10.9. 25 Villages & camps to be sampled were calculated using

Where: n=number of children with SAM to be sampled, average village population within Kismayo town (accessible area) 57026, percentage population of children 6-59 months 20%, prevalence of SAM of 2.7%27. A precision of 10% was utilized. 4.2 Quantitative sampling framework: A list of all the villages and camps within the accessible Kismayo area (town) was used to select villages and camps for the wide area survey. Since the small survey did not show one area having a higher coverage than the other does, a comprehensive list of villages and camps was used to determine sample villages. Twenty-five villages were randomly selected from a list the total 111 villages/camps in Kismayo district (urban) served by the OTP sites.

4.2.1 Case finding methodology Being that stage 2 revealed that the coverage was already low it was suggested by the program staff that a ‘door to door ‘case finding approach would be more beneficial in order to capture all uncovered cases in the sampled villages. The mobile phone contact of all the caregivers of the uncovered children, both current SAM and recovering SAM cases, were taken for follow up and referral. Children who fit into the SC protocol were referred to the SC during the exercise.

4.2.2 Quantitative data results: Children who were screened in the villages/camps during the wide area survey were categorized as SAM in OTP, SAM not in OTP and recovering SAM in program (Annex 4). In calculation of the single coverage estimate, the calculated recovering-out (Rout)28 was used (Table 7).

Table 7: Results of the wide area survey Indicator Number A SAM in OTP 59 B SAM not in OTP 97 C Total current SAM cases 156 D Recovering in program 48 E Recovering out (calculated) 25 F Total cases 22929

The single coverage estimate was 46.8% (40.8% - 53.1) where z = 0.05, p = 0.959.The plot has a coherent prior and likelihood (z value near 0 and p value almost 1) to mean that the barriers and boosters achieved during the assessment were understood and representative of the situation on the ground (Figure 7). The coverage estimate was however below any of the Sphere standards for program coverage (IDP 90%, Urban 70%, rural 50%). This instigates that the barriers have a lot of effect on the Kismayo coverage and need to be addressed with urgency.

26 MoP Kismayo 2016 IDP population estimates; Village elders and program coordinator for host estimates 27 FSNAU Post Deyr 2015/2016 national prevalence 28 www.ennonline.net/fex/49/singlecoverage 29 Single coverage estimation(A+D)/F; Point coverage (A/C); Period coverage (A+D)/(C+D) 19

Figure 8: Beta binomial conjugate analysis of single coverage estimate

4.2.3 Reasons for SAM non-attendance There were a variety of reasons given by caregivers of SAM cases not in OTP most of whom said they had been at work or they had no choice to care for their children (Figure 8).

Figure 9: Caregiver reasons for not taking their malnourished children for OTP Twenty percent of the caregivers interviewed did not know that their children were malnourished. Some of those children had been discharged to home and now relapsed (29%) while some had been ill (29%) in the days leading to the assessment. The other popular responses inlcuded recent arrivals from neighbouring areas (Gosha) or lack of knowledge of the program. There were cases of program rejection due to the preferred admission criteria (W/H Z- scores) that led some to fear going through the same situation. For afew however, they just preferred treatment from the area’s traditional healer (Figure 8).

RECOMMENDATIONS Section A Section B Section C Section D

Barrier Recommendations (Actions to be taken) Who is Indicators of progress Timelines What activities can be done to responsible reduce the effect or weight of barrier/ way forward? RUTF as food Sensitization meetings in the Number of sensitization meetings MoH and community on malnutrition and its implementing Quarterly treatment by MoH with village elders, Number of shops that sell RUTF organizations and other community groups (including shopkeepers), influential individuals in the community Beneficiaries to bring back the empty Number of beneficiaries who bring back Program staff Weekly sachets during their visits before being the empty sachets and OTP team given more ration leaders Sensitize caregiver groups on income Number of caregiver groups with income Monthly generating activities in-order to generating activities counter sale of RUTF Community screening Admission into the program to Number of MUAC OTP admissions UNICEF and Monthly by MUAC while consider all the admission criteria Implementing admission by W/H Z including MUAC organization scores only Caregivers to be informed of the Number of caregivers who acknowledge Implementing Weekly admission procedures without promise the process after referral program of OTP admission OTP sites not integrated OTP sites to scale up the services Number of OTP sites with medical UNICEF and Quarterly with other medical offered to include some medical treatment implementing services treatment services organization and MoH Weak OTP-SFP Have a line of feedback to ensure the Number of confirmed referrals by OTP Implementing Monthly interface OTP-TSFP referrals reach the sites and TSFP sites Organizations they are sent to for continuity of treatment Increase number of TSFP sites Number of TSFP sites MoH, Donors Quarterly 21

and implementing organizations Scale up some OTP sites to manage Number of OTP sites managing SAM UNICEF, other Quarterly malnutrition cases until the discharge cases to cured from MAM donors and criteria for MAM implementing organizations Limited information Have agenda for program information Number of coordination meetings Cluster Quarterly sharing among partners sharing with partners through the coordinator, coordination meetings Partners can identify services and MoH operational sites in the community they serve Stigma Community sensitization meetings on Number of community sensitization MoH and Quarterly causes of childhood malnutrition meetings on causes of childhood Implementing especially among fathers malnutrition among fathers organization

Number of late OTP admission Late treatment seeking Sensitization meetings with Number on early treatment seeking MoH and Quarterly community members on need for community sensitization meetings Implementing medical treatment organizations Number of SC admissions

Number MUAC at admission into OTP Increase the number of medical Number of medical treatment centers MoH, Quarterly treatment centers offering affordable UNICEF, other services donors and implementing organizations Stock-out of RUTF Increase the buffer ration to cater for Number of RUTF stock outs Implementing challenged transportation (Proper organizations Caseload calculation and and UNICEF communication should be put in place) and other RUTF supply donors Clear and timely forecasting and Number of RUTF stockouts versus Implementing Quarterly 22

communication on RUTF needs timely forecasting and communication organizations on RUTF needs and UNICEF and other RUTF supply donors Influx of families and Increase active case finding during the Number of large scale active case finding MoH, Every other IDPs from neighboring months of population influx conducted Implementing month areas organization

Annexes Annex 1: List of people trained during SQUEAC Assessment NAME POSITION ORGANISATION Abdimalik Barkhatie Male Enumerator Government State house Abdinasir Hassan Male Supervisor SAF UK Abdi Omar mohammed Male Enumerator Ministry of Agriculture Hassan Mohammed Abdulahi Male Supervisor SAF UK Sugulle Abdi Hassan Male Enumerator SAF UK Fartuun Mohamud Mohamed Female Enumerator PAC Sahan Mohamed Ali Female Enumerator Somali Aid Dekow Yussuf Maalim Male Enumerator WRRS Abdullahi Ibrahim Abdi Male Enumerator Host community member Hodan Ali Farah Female Enumerator ICRC Abdikadar Mohamed Ali Male Enumerator Ministry of Health JSS Abshiro Abdinor Yasim Female Enumerator Himillo Foundation Everlyne Adhiambo Female Program Coordinator SAF UK Abdikarim Ali Abdi Male Program manager SAF UK Annex 2: Training chronogram

Time Activity

9:30- 10.45 - Enumerator terms of engagement - Introduction to SQUEAC 10.45- 11.10 Tea break 10.10 – 1.00 - Stage 1 Quantitative data - Brainstorm on seasonal calendar considering trends in program data 1.00-2.00 Lunch break 2.00-5.00 - Qualitative data collection tools - Individual team formation Discussion forum for translation of tools to Somali 24

Annex 3: Kismayo OTP operational days

Zone IDP/Host Urban/Rural OTP Site Operational days Dabyanbo Fixed Host 5 days Urban Farjano Dalxiska Fixed IDP 5 days Urban Mudul Mobile IDP SAT Urban Fanole Marina Mobile Host WED Urban Warshada Hargaha Mobile IDP SUN Urban Gulwade Adey Marogto Mobile Host MON, TUE Urban Maryan Khamis Mobile Host WED Urban Shaqaalaha Kam Hilac Mobile Host SAT, SUN, MON, TUE Urban Kam Jibirti Mobile Host TUE, WED Urban Alanley Dowdhanan Mobile Host SAT, SUN, MON Urban

Annex 4: Wide area survey quantitative data

Village or camp SAM IN OTP (MUAC SAM NOT in OTP (MUAC Recovering SAM IN Program <11.5cm or oedema) < 11.5cm or oedema) MUAC ≥11.5cm – 12.5cm Waamo 2 0 2 0 Talax 4 2 1 Bader 10 14 3

Baqdaad 2 3 8 Haamdi 0 3 3 Tawakal 1 1 3 0 Nageye 0 1 0 Muminomarketi 0 6 0

Marino 5 4 11 Dhumasa 2 8 1 Towfiq 10 6 2 Aruriyow 1 1 0 Ahmed Binuhambal 0 5 0 1 August 0 2 0 Wardheer 0 4 5 Baas 3 0 3 0 Halgan 1 2 1 Najii Mahamuse 10 3 10 Lafola 0 0 0 Badar 10 14 3 Dhagahtor 0 3 0 Galey camp 0 2 0 Yaman 3 4 0 Iftin 0 2 0 Total 59 97 48

Annex 5: Health facilities Kismayo

Facility name Facility type Management Urban/rural Alanley MCH Mother Child Health/Health Center Muslim Aid Urban APD MCH Mother Child Health/Health Center APD Urban Buula-Ablika MCH Mother Child Health/Health Center ARC Urban Ex- Midwifery MCH Mother Child Health/Health Center Somali Aid Urban Fanolle MCH Mother Child Health/Health Center Muslim Aid Urban Farjano MCH Mother Child Health/Health Center SRCS Urban Gulwade MCH Mother Child Health/Health Center MoH Rural Kismayo General Hospital Hospital ICRC/PAC Urban Kismayo General Hospital MCH/OPD Mother Child Health/Health Center Somali Aid Urban Muslim Hand MCH Mother Child Health/Health Center Muslim Hand Urban Muslim TB Centre TB Center Muslim Aid Urban Siinay MCH (Mahfalka) yagleel Mother Child Health/Health Center Non functional Urban Siinay MCH ARC Mother Child Health/Health Center ARC Urban SWACEDA Hospital Hospital SWACEDA Urban Waaberi MCH Mother Child Health/Health Center MoH Urban

Annex 6: Concept map for Kismayo

Drawn using epigram software

Annex 7: Kismayo Urban population estimates Kismayo Urban population

Population Host/ IDP Source

Farjano zone 4 sections 10,2587 Host population

Fanole zone 5 sections 95,287 Host population Kismayo village elders and Alanley zone 6 sections 58,836 Host population implementing organization (SAF UK)

Shaqaalaha 5 sections 88,893 Host population zone

IDPs 91 camps 30,690 IDP population Ministry of planning Jubaland state, Somalia

Annex 8: Map of Somalia showing position of Kismayo District

Source: Adapted and modified from http://www.fsnau.org/products/maps/administrative-maps